mlsanityData to 7 Oct 2026

Learning to Drive via Asymmetric Self-Play

arXiv 2409.18218 ↗First posted 26 Sep 2024cs.RO cs.CV cs.LGcited by 11

On the map: End-to-end autonomous driving → Imitation-learning end-to-end planners

Abstract (arXiv)

Large-scale data is crucial for learning realistic and capable driving policies. However, it can be impractical to rely on scaling datasets with real data alone. The majority of driving data is uninteresting, and deliberately collecting new long-tail scenarios is expensive and unsafe. We propose asymmetric self-play to scale beyond real data with additional challenging, solvable, and realistic synthetic scenarios. Our approach pairs a teacher that learns to generate scenarios it can solve but the student cannot, with a student that learns to solve them. When applied to traffic simulation, we learn realistic policies with significantly fewer collisions in both nominal and long-tail scenarios. Our policies further zero-shot transfer to generate training data for end-to-end autonomy, significantly outperforming state-of-the-art adversarial approaches, or using real data alone. For more information, visit https://waabi.ai/selfplay .

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Placed in Imitation-learning end-to-end planners, one of 9 directions in End-to-end autonomous driving. Builds on 72 of its 92 references that are in the map. First posted on arXiv 26 Sep 2024.

Builds on

72 of 92 references.

Closest ancestors · 6

Same direction, newest first

  1. 2023Waymax: An Accelerated, Data-Driven Simulator for Large-Scale Autonomous Driving Research12 Oct
  2. 2023Rethinking Closed-loop Training for Autonomous Driving27 Jun
  3. 2023TrafficBots: Towards World Models for Autonomous Driving Simulation and Motion Prediction7 Mar
  4. 2022Imitation Is Not Enough: Robustifying Imitation with Reinforcement Learning for Challenging Driving Scenarios21 Dec
  5. 2022Embedding Synthetic Off-Policy Experience for Autonomous Driving via Zero-Shot Curricula2 Dec
  6. 2022KING: Generating Safety-Critical Driving Scenarios for Robust Imitation via Kinematics Gradients28 Apr

Other references · 66

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  4. 2017VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detectioncited by 1,323
  5. 2019Dota 2 with Large Scale Deep Reinforcement Learningcited by 1,084
  6. 2016End to End Learning for Self-Driving Carscited by 1,000
  7. 2023Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecastingcited by 701
  8. 2021Large Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Datasetcited by 532
  9. 2017End-to-end Driving via Conditional Imitation Learningcited by 479
  10. 2018ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worstcited by 391
  11. 2020Learning Lane Graph Representations for Motion Forecastingcited by 380
  12. 2021MetaDrive: Composing Diverse Driving Scenarios for Generalizable Reinforcement Learningcited by 281
38 more
  1. 2018Learning to Drive in a Daycited by 266
  2. 2021End-to-end Interpretable Neural Motion Plannercited by 240
  3. 2020Fast and Furious: Real Time End-to-End 3D Detection, Tracking and Motion Forecasting with a Single Convolutional Netcited by 228
  4. 2021MP3: A Unified Model to Map, Perceive, Predict and Plancited by 211
  5. 2021TrafficSim: Learning to Simulate Realistic Multi-Agent Behaviorscited by 192
  6. 2022Guided Conditional Diffusion for Controllable Traffic Simulationcited by 189
  7. 2021Scene Transformer: A unified architecture for predicting multiple agent trajectoriescited by 165
  8. 2023MotionLM: Multi-Agent Motion Forecasting as Language Modelingcited by 154
  9. 2020PnPNet: End-to-End Perception and Prediction with Tracking in the Loopcited by 143
  10. 2017Intrinsic Motivation and Automatic Curricula via Asymmetric Self-Playcited by 139
  11. 2020LiDARsim: Realistic LiDAR Simulation by Leveraging the Real Worldcited by 137
  12. 2021Generating Useful Accident-Prone Driving Scenarios via a Learned Traffic Priorcited by 136
  13. 2021AdvSim: Generating Safety-Critical Scenarios for Self-Driving Vehiclescited by 132
  14. 2023Language-Guided Traffic Simulation via Scene-Level Diffusioncited by 110
  15. 2022On Adversarial Robustness of Trajectory Prediction for Autonomous Vehiclescited by 89
  16. 2023The Waymo Open Sim Agents Challengecited by 86
  17. 2022BITS: Bi-level Imitation for Traffic Simulationcited by 80
  18. 2021SimNet: Learning Reactive Self-driving Simulations from Real-world Observationscited by 77
  19. 2018Scalable End-to-End Autonomous Vehicle Testing via Rare-event Simulationcited by 76
  20. 2020Learning to Collide: An Adaptive Safety-Critical Scenarios Generating Methodcited by 74
  21. 2021Asymmetric self-play for automatic goal discovery in robotic manipulationcited by 71
  22. 2019Adaptive Stress Testing for Autonomous Vehiclescited by 70
  23. 2022GoRela: Go Relative for Viewpoint-Invariant Motion Forecastingcited by 68
  24. 2022Occupancy Flow Fields for Motion Forecasting in Autonomous Drivingcited by 62
  25. 2022Symphony: Learning Realistic and Diverse Agents for Autonomous Driving Simulationcited by 61
  26. 2022Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real worldcited by 56
  27. 2019Jointly Learnable Behavior and Trajectory Planning for Self-Driving Vehiclescited by 51
  28. 2019Model-Predictive Policy Learning with Uncertainty Regularization for Driving in Dense Trafficcited by 50
  29. 2018Multi-Agent Imitation Learning for Driving Simulationcited by 49
  30. 2022Mastering the Game of No-Press Diplomacy via Human-Regularized Reinforcement Learning and Planningcited by 47
  31. 2023RealGen: Retrieval Augmented Generation for Controllable Traffic Scenarioscited by 46
  32. 2018Scenarios for Development, Test and Validation of Automated Vehiclescited by 45
  33. 2020Adversarial Evaluation of Autonomous Vehicles in Lane-Change Scenarioscited by 42
  34. 2020Scenic: A Language for Scenario Specification and Data Generationcited by 39
  35. 2019Failure-Scenario Maker for Rule-Based Agent using Multi-agent Adversarial Reinforcement Learning and its Application to Autonomous Drivingcited by 36
  36. 2020Neural Bridge Sampling for Evaluating Safety-Critical Autonomous Systemscited by 34
  37. 2019Adaptive Stress Testing with Reward Augmentation for Autonomous Vehicle Validationcited by 33
  38. 2021Learning to Simulate Self-Driven Particles System with Coordinated Policy Optimizationcited by 33

50 most cited of 66 All 66 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 6; 2026: 5.

11 papers.

  1. 2025SimScale: Learning to Drive via Real-World Simulation at Scalecited by 34
  2. 2025Building reliable sim driving agents by scaling self-playcited by 20
  3. 2025SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Modelcited by 20
  4. 2025Reinforced Refinement with Self-Aware Expansion for End-to-End Autonomous Drivingcited by 19
  5. 2025SPACeR: Self-Play Anchoring with Centralized Reference Modelscited by 7
  6. 2026Learning to Drive in New Cities Without Human Demonstrationscited by 6
  7. 2026World Engine: Towards the Era of Post-Training for Autonomous Drivingcited by 6
  8. 2026Scaling Self-Play for End-to-End Drivingcited by 4
  9. 2025Heterogeneous Adversarial Play in Interactive Environmentscited by 3
  10. 2026Pictura: Perspective-View Self-Play at Scale for Drivingcited by 2
  11. 2026Conditional Flow-VAE for Safety-Critical Traffic Scenario Generationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026TerraTransfer: Learning End-to-End Driving Policies Without Expert Demonstrations
  2. 2021Urban Driver: Learning to Drive from Real-world Demonstrations Using Policy Gradients
  3. 2026TerraZero: Procedural Driving Simulation for Zero-Demonstration Self-Play at Scale
  4. 2026Human-like autonomy emerges from self-play and a pinch of human data
  5. 2024Data Scaling Laws for Imitation Learning-Based End-to-End Autonomous Driving
  6. 2025Learning to Drive from a World Model
  7. 2023Scaling Planning for Automated Driving using Simplistic Synthetic Data
  8. 2021Learning Interactive Driving Policies via Data-driven Simulation